[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: Re:overriding the schema at runtime..

• From: Michael Glavassevich <mrglavas@c...>
• To: "Pete Cordell" <petexmldev@c...>
• Date: Mon, 7 Apr 2008 23:29:21 -0400

"Pete Cordell" <petexmldev@c...> wrote on 04/07/2008 02:45:37 PM:

> There's no real way to override processContents="strict".  I think you
just
> have to accept that the schema contains a bug in it.  As it happens using

> processContents = 'skip' is also the right way to go for including
elements
> with namespaces.
>
> One other thing, I've seen some reports of processors having problems
with
> schemas that specify specific, but high values for maxOccurs (e.g. 50 as
> opposed to unbounded).  This is because they try to build a state machine

> with as many states as maxOccurs dictates.

Worth noting that some of these processors have since changed the way that
they check maxOccurs. For instance Xerces-J used to always build up a state
machine proportional to the size of maxOccurs. Now it represents these with
a single state for simple cases (e.g. maxOccurs > 1 on
xs:element/xs:wildcard) which loops with a counter to track how many
instances there are.

> I'm not sure at the point where the problems start arising, so maybe 50
> is OK.  Just thought I'd mention it as something to bear in mind!
>
> HTH,
>
> Pete Cordell
> Codalogic
> For XML C++ data binding visit http://www.codalogic.com/lmx/

Thanks.

Michael Glavassevich
XML Parser Development
IBM Toronto Lab
E-mail: mrglavas@c...
E-mail: mrglavas@a...

• References:
  • Re:overriding the schema at runtime..
    • From: "Pete Cordell" <petexmldev@c...>